Debates about different consensus mechanisms of blockchains belong to the most persistent chatters in the crypto hemisphere since its formation. Blockchains depend on a Proof-of-Something in order to prove the dedication of a scarce resource to the protocol. In Proof-of-Work mining hardware and the consumption of energy are dedicated whereas in Proof-of-Stake it is the end-protocol asset itself (in this case: ETH). In contrast to a just hypothetical Proof-of-Flowers, the dedication of both work and stake are easy to prove to a protocol. This is why they serve as prominent representatives of crypto-economic consensus mechanisms.
First off and to eradicate a popular misconception, with The Merge (referring to the point in the near future at which the Ethereum blockchain will switch consensus mechanism from PoS to PoW), transaction fees will NOT become significantly cheaper. Why is that?
PoS introduces a reduction in time between blocks from 13-14 seconds to 12 seconds and leads to less variance in timing between blocks. This implies just a marginal increase in transaction throughput - measured in transactions per second (TPS) - and obviously represents everything else but the main driver for increased scalability.
However, PoS opens the door for more sophisticated consensus mechanism features enabling to scale the blockchain - such as sharding.
On a very high level sharding means slicing up the decentralised database and allowing for randomly sampling consensus participants (i.e. validators) in a secure way to exercise distributive work in different subsets of the protocol.
This feature is very difficult to implement in PoW, because consensus participants (i.e. miners) exist outside the protocol and cannot be easily coordinated in a way to fulfil complex tasks such as sharding.
While roll-ups serve for scaling by outsourcing the execution of transactions, a sharded blockchain will allow for more data to be processed on-chain (e.g. validity or fraud proofs of off-chain transaction execution). This is why PoS represents a key cornerstone for Ethereum’s rollup-centric roadmap and enables modularizing the blockchain stack even further in order to meet higher blockspace demand as global adoption advances.
Improved user experience is a rarely mentioned advantage of PoS as it introduces stronger economic finality guarantees than PoW - a highly desirable blockchain property. This means validators are able to guarantee that block history cannot be reverted up until a certain checkpoint. The lower the time-to-finality, the earlier a user or businesses receive confirmation that their transaction has made it on-chain and cannot be reverted.
In PoS users will wait on average 6 seconds for the next block, and at the longest 12 seconds, while benefiting from more than hundred confirmations per block which is why they receive security assurances much faster than today. This applies for example when centralized exchanges urge their users to wait for xyz more confirmations in order for their deposit or withdrawal to be processed.
PoS also enshrines light clients (or light nodes) as pivotal network infrastructure in order to foster decentralisation and improve user experience. Think of them as functionally specialized and much more lightweight, bandwidth-reduced and cost-efficient consensus participants running on mobile phones and/or as browser plug-ins. Users can attach their wallets to one or more light clients and trustlessly validate the latest balance of an account (among other things), hence not depending on centralized node infrastructure any more.
On another note, Ethereum’s sophisticated merge process design likewise offers remarkable user experience given the fact that the task at hand literally equals changing wheels while driving at full speed. From a user’s perspective the switch from PoW to PoS means:
As blockchain technology gains broader adoption, more and more value is settled on them. In 2021 Ethereum settled >11 trillion USD. The more value stored on-chain, the more lucrative an attack on a blockchain appears. Not everybody benefits from disruptive and decentralized technology. Blockchains do have powerful opponents, which is why Ethereum protocol developers are particularly keen on securing the network and literally strive for WW3-resistance.
One of the main advantages of PoS regarding its superior security properties is its less centralizing nature with respect to its own crypto-economic asset, ETH, representing the dedicated resource for consensus participation:
PoS lacks economies of scale. The barrier to entry staking is as low as 0.01 ETH. On the contrary, in PoW larger scale actors benefit from favourable deals with chip manufacturers (and energy providers), while hobbyists often suffer from being forced to mine on older, less efficient hardware to prove their work.
Secondly, staking rewards as returns for performing validator duties are paid in ETH on what is at stake - which is also ETH. The convergence of rewards and the underlying crypto-economic asset to just ETH
provides a much more efficient and purer function of the return on capital employed. By contrast, PoW rewards paid in ETH currently flow back to consensus participants already operating at scale (mining farms) thereby centralizing the asset flow even further.
It is therefore reasonable to assess that in PoW the rich get richer.
ETH the asset is highly liquid and very easy to obtain on a large amount of open markets all over the world, consequently lowering the barrier to participate in staking and thus potentially decentralising the network even more as opposed to PoW. Interestingly enough, PoS even paves the way for ETH as crypto-economic asset to be hypothecated and used as a productive asset in DeFi.
PoS also makes a blockchain more secure against attacks, because the capital at risk for consensus participants is by design higher than in PoW. An appropriate measure for this is the security margin of a blockchain which refers to the consensus participant’s amount of crypto-economic capital at risk.
In PoW miners invest in mining hardware upfront and receive miner revenue (= block reward + transaction fees) as returns, or - in case of misbehaving to the protocol rules - do not receive the rewards. Their (physical) mining farm, though, is practically never at risk of being expropriated. Contrary to this, for a validator in PoS only ETH the asset is at risk. It serves as the dedicated and bonded in-protocol resource for consensus participation.
In addition, however, PoS allows for penalizing a consensus participant in case of being idle or offline. In an extreme case of severe misbehaviour or attack it even allows for slashing a validator, setting up to 100% of its stake at risk of being lost and removed from ETH supply. The protocol itself literally is in command of identifying the attacking validator(s) and burning down their (virtual) mining farm, thereby significantly increasing the security margin of the chain.
On a rather technical note, PoS moreover offers better options for emergency recovery in case of an attack. In PoW an attacker with 51% of hash-rate basically gains ‘god-mode’ and can arbitrarily re-organize the chain and double-spend coins over and over again. Vitalik once framed this spawncamping in perpetuity. The protocol has no chance to recover, because mining hardware and energy as attacking resources exist outside the protocol and thus out of reach. ETC and BSV provide cautionary examples suffering from such attacks and in the following turned out practically unusable.
PoS recovery, by contrast, would mean burning the attacking stake (i.e. bonded capital in the form of ETH) of the now identifiable validator(s) so that they only have one costly shot at rewriting chain history at all. Any additional attempt a) is disincentivised by the limited overall supply of ETH and b) becomes even more expensive as a subsequent shift in supply/demand mechanics moves the price of ETH to a new unfavourable equilibrium for any attacker.
As an ultima ratio PoS would allow for social intervention, meaning the community could potentially socially coordinate and fork off by forking out evil actors’ assets, because they are identifiable. This would of course equal an extremely theoretical, unlikely and highly undesirable event. Nevertheless it represents a powerful, disincentivising threat and is at least a possible option compared to PoW where attacking miners cannot be socially forked out as they are unidentifiable.
Increased sustainability with PoS is twofold: ecologically and economically.
With the absence of hardware intensive and energy consuming mining activities, PoS ensures a decline in energy consumption of more than 99.9%. The underlying assumption is based on the current Ethereum energy consumption of around 110.000 GWh per year, which lands the network an unflattering place in the Top 30 list of countries.
Let’s assume an average laptop for running a staking rig consumes ~60 watts per hour and runs 24/7 a year. If currently around 10k Ethereum nodes worldwide each consumed a corresponding 525 KWh per year, this would sum up to about 5 GWh - the equivalent of a smaller town in an industrialized country. Entering PoS the nodes will remain, but
110.000 GWh consumed for mining will vanish. Even if, fortunately, the decentralization of the network increases further and thus the amount of nodes multiplies, the reduction in energy consumption remains immense.
All this is enabled by eliminating the necessity for providing dedication of work to secure the network.
The dedication of scarce resources to participate in consensus is not energy consumption any more, it is ETH itself.
Moreover, it is worth noting that PoS escapes the undesirable vicious circle in PoW where the higher the price of the chain’s native coin, the higher the incentive to spend more energy to mine it (scaling proportionally).
As shown above, due to the lack of costly resource and energy consumption PoS allows for reducing the issuance of new coins paid to incentivize consensus participants. In fact, it enables spending less for the security of the blockchain by paying a minimum viable issuance and thus economically sustaining the protocol as well. In conjunction with a sophisticated fee market design - instantiated as part of EIP-1559 in 2021 - it is even possible to experience phases of net-negative issuance, bolstering the tokenomics of ETH unlike any asset before.
Proof-of-Stake is arguably more complex and nuanced than Proof-of-Work. Nevertheless Proof-of-Work has always been considered a bridging technology for the Ethereum blockchain. Significant gains in security and efficiency make Proof-of-Stake the consensus mechanism of choice for the future.
Consequently, the doom of Proof-of-Work is near. It is time to move onward.
The inspiration for this post stems from a variety of research initiatives combined with multi-year development efforts. If you sympathise with sustainable Ethereum protocol development check out the Protocol Guild - a public goods funding mechanism to support this cause.